In the history of commercial deployment of fiber networks, the dominating tendency was to minimize the spectral width of transmitted signals. There have been at least two motivations to pursue spectral minimization: to reduce signal dispersion within transmission line and to simplify the equipment for signal transmission in Non-Return-to-Zero (NRZ) format.
A majority of contemporary WDM transmission systems operate at a data rate of 2.5 Gb/s (OC-48) and utilize standard single-mode fiber (SMF) with chromatic dispersion D about 17 ps/nm*km. In these systems, the transmission distance is limited to about 600 km and does not require dispersion compensation. For WDM systems of higher bit rates (10 and 40 Gbit/s) and similar reach, dispersion compensation is required. Deployment of dispersion compensating means such as dispersion compensating fibers or fiber gratings eliminates unwelcome linear dispersion impairment [A. H. Gnauck and R. M. Jopson, “Dispersion Compensation for Optical Fiber Systems”, Chapter 7 in Optical Fiber Communication Systems, Vol. IIIA, Ed. I. P. Kaminow and T. L. Koch, Academic Press, San Diego, 1997].
The performance of dispersion compensated systems with periodic optical amplification is limited by accumulation of spontaneous emission noise and fiber non-linearity. To optimize the NRZ-format system performance, a fine balance has to be achieved between maximizing optical signal-to-noise ratio and minimizing nonlinear effects such as self-phase modulation (SPM), cross-phase modulation (XPM), and four-wave-mixing (FWM). Conventional power per individual channel at the input of each fiber span is about 1 mW (0 dBm), resulting in a maximum propagation distance of several hundred kilometers. For single channel transmission and for WDM multi-channel transmission in non-dispersion-shifted fibers, SPM is the dominating nonlinear effect. An optical signal propagating in the fiber experiences an intensity dependent phase shift φNL(t,z) given by [G. P. Agrawal, Nonlinear Fiber Optics, Chapter 4, Academic Press, San Diego, 1989]:φNL(t,z)=γP(t)z,  (1) where γ is the nonlinear coefficient of the fiber, P is the optical power, and z is the effective propagation length. Because the signal intensity is time-dependent, new spectral (frequency) components are produced:Δω=−dφNL/dt.  (2) When coupled with dispersion, these new spectral components interfere with the original components and distort the temporal shape of the signal. The NRZ data stream is a complicated temporal pattern, in which each individual bit is distorted differently than other bits and therefore cannot be globally restored. As a result, NRZ-format transmission systems are designed to work in the linear regime with a small nonlinear phase shift, φNL(t,z)<<1. In general, utilization of NRZ format and minimization of the spectral content limits further development of transmission technology.
Several approaches were developed to overcome the non-linear distortion problem, all of them based on transmitting short optical pulses, or using return-to-zero (RZ) format. The important advantage of RZ format is that the distortion caused by SPM is pattern-independent. In RZ format, every single bit representing 1 is a single pulse identical to other pulses. If a single pulse is compensated for the SPM distortion, then the whole data stream is compensated automatically.
In retrospect, the first method of nonlinear distortion compensation was the propagation of optical solitons [L. F. Mollenauer, J. P. Gordon, and P. V. Mamyshev, “Solitons in High Bit-Rate, Long-Distance Transmission”, Chapter 12 in Optical Fiber Communication Systems, Vol. IIIA, Ed. I. P. Kaminow and T. L. Koch, Academic Press, San Diego, 1997]. To create an optical soliton, a Fourier-transform limited optical pulse having maximum power of several dBm is propagated in optical fiber with small positive dispersion. The general idea of soliton propagation is that the pulse shape is affected by chromatic dispersion and non-linearity in a way that these two factors counterbalance each other. For a certain range of pulse duration and optical power the pulse shape is kept constant along the fiber. The trend of the pulse to expand due to linear dispersion is compensated by contraction due to self-phase modulation so that the pulse is kept intact during the propagation. In other words, the combined effect of linear dispersion and SPM on the propagating pulses is continuously compensated at every point in the fiber. Using soliton transmission technology, Fourier transform limited light pulses of about 10-30 ps duration and peak power of several dBm may be propagated along fiber spans of many thousands of kilometers without significant shape degradation.
The balance between linear and nonlinear contributions, however, could be maintained only within certain ranges of pulse power and dispersion (D<˜1 ps/nm*km). The low dispersion requirement limits the applicability of soliton transmission to special types of fiber such as dispersion-shifted fibers (DSF), where the zero wavelengths dispersion is 1.5 micron. The power range requirement means that in-line optical amplifiers must be spaced much closer than the customary span length for commercial terrestrial long-haul networks. Further complication is caused by soliton timing jitter (known as the Gordon-Haus effect) induced by the soliton coupling with the accumulated amplifier noise. Special filtering schemes were developed to reduce the timing jitter and allow for extra long propagation of signals. However, these schemes are generally too complicated to deploy in commercial transmission systems. As a result, the soliton transmission though studied for almost 20 years has not found commercial applications.
In the last few years, a new approach for RZ transmission has been developed called dispersion-managed solitons, or quasi-solitons (M. Suzuki et al, Electronics Lett., Vol. 31, p. 2027, 1995; J. H. B. Nijhof et al, Opt. Lett., Vol 23, p. 1674, 1998; F. Favre et al, Jour. Lightwave Tech., Vol 17, p. 1032, 1999). According to this approach, the nonlinear dispersion is compensated within each span of fibers in a periodically optically amplified fiber transmission line, instead of being continuously compensated at each point of the fiber as in standard soliton transmission. It is based on a basic phenomenon of propagating of linearly frequency chirped pulses in nonlinear media: when the linear chirp and non-linearity-induced chirp of the pulse are in the same direction the spectral bandwidth of the pulse increases; when they are in the opposite directions the spectral bandwidth decreases. If the linear dispersion of the media periodically changes sign in a properly designed way, the spectral content of the propagating pulse and its shape would also vary periodically. By using appropriate pre-dispersion and arrangement of fibers of positive and negative dispersions, commonly referred to as “dispersion map”, one could periodically reconstruct the original pulses, both in terms of duration and spectral bandwidth, at each optical amplifier site along the transmission line. However, to achieve this indefinite oscillatory mode, the dispersion management has to be very precise (have extremely small tolerances) which is difficult to implement in practice.
Therefore, the telecommunication industry is in need of new methods and systems for transmitting signals via multiple spans of optical fiber without considerable linear or non-linear distortions.